MyFood Dataset

MyFood Dataset is an image database for segmenting images of Brazilian foods. Composed of 9 classes: rice, beans, boiled egg, fried egg, pasta, salad, roasted meat, apple and chicken breast. With an average of 125 images per class and a total of 1250 images, with a ratio of 60-20-20 for the training, validation and testing sets, respectively.

All images have their respective food segmentation masks, according to the folder structure:

Note: mask_1.png corresponds to the mask of img_1.png, and so on.

MyFoodDataset/

          - train/

                 img_1.png

                 img_2.png

                 …

          - train_ann/

                 mask_1.png

                 mask_2.png

                 …

           - val/

                 img_1.png

                 img_2.png

                 …

           - val_ann/

                mask_1.png

                mask_2.png

                …

            - test/

                img_1.png

                img_2.png

                …

            - test_ann/

                mask_1.png

                mask_2.png

                …